Every restaurant operator knows the recipe for the perfect schedule: get the right people on the right shift at the right time. But when it comes down to it, that’s much easier said than done.
Restaurant operators have been juggling various approaches to scheduling, hoping to achieve the highest possible level of accuracy. While the methods for scheduling have evolved, each iteration comes with its own set of challenges. Below, we outline the evolution through the three most common approaches to scheduling—from relying on hunches to relying on science—and share insight into the more accurate fourth (pun intended!) approach: demand forecasting.
Scheduling based on percentage of weekly budget
The traditional approach to scheduling—basing staffing on the weekly budget—means mangers allocate labor across the week based on their allowed labor spend, based on expected sales. What does that look like?
Managers have a fixed amount of money to spend on staffing throughout the week and assign team members to set shifts. With little thought to staggering staff shifts or varying activity levels during each hour in a given shift, the resulting day has periods of time that are overstaffed, while peaks around certain popular times feel thin.
Usually, too many employees are on the schedule for the beginning of the week. The error compounds, and in order to stay on budget, managers are forced to cut shifts at the end of the week. Unfortunately, these also tend to be the busiest, so you run the risk of impact to the guest experience. On the flip side, if managers assume a lighter front-end of the week, an unexpected change in weather or a holiday could drive an unanticipated amount of traffic to the restaurant.
When scheduling to a budget, managers also run the risk of overstaffing on slower weeks (or deliberately overstaffing to avoid budgets being cut). A slow, boring shift means employees receive lower tips and have less to do, leading to employees who are checked out or distracted on the job. Guests may have a hard time getting their server’s attention or, conversely, may feel that staff is overbearing and over-attentive. Meanwhile, every over-staffed hour is denting your bottom line.
Many restaurants stick with this traditional approach to scheduling because it does help them stay within their set weekly budgets, which is, of course, important for regulating spend. The impact, however, is on team members (who take the brunt of the stress), the loss to potential revenue (as key selling opportunities are missed and future recurring revenue is less likely with unhappy customers) and on the cost of employee turnover (when employees aren’t making enough in tips or aren’t scheduled enough, they leave to find other work). While it may seem to work in the weekly short-term, the long-term impact to your bottom line and growth potential is real.
Sales Per Labor Hour (SPLH)
SPLH forecasting works by attributing a value to each labor hour. With this figure, the sales revenue forecast can then be divided by it to decide how many employees are needed. While SPLH targets differ from role to role, a rough example would be that if the total sales for the week are $2,000 and your target SPLH is $50, then your manager has 40 hours to work with. As you may guess, the main challenge with scheduling by SPLH is that the workload or staff-impact differs greatly depending on what is actually ordered. $150 can be made serving a single bottle of champagne or running 50 separate $3 soft drinks to different tables, and the time spent by a given employee differs for each scenario. The people you need differs, too, for different activities: serving a beer requires different training from shaking up a cocktail, for instance.
The other limitation is that the forecasted spend does not account for the tasks that do not generate revenue but require staff attention and time. For example, what percent of a given shift is taken up by prep work or restocking? Often, these hours are not accounted for when forecasting by SPLH.
SPLH is undoubtedly a useful metric as it allows operators to benchmark employees and identify areas for training and improvement. But using SPLH for forecasting doesn’t provide the appropriate level of accuracy for your business.
Scheduling based on covers
Some restaurants have now turned to scheduling based on covers—the number of seats filled on a given shift—as a way to more accurately predict staffing needs. This does bring in some science to the equation, and covers are good metric and can get operators a step closer to their goal of accuracy. But this approach still has major limitations when it comes to forecasting the future and analyzing past results.
The main challenge here is that scheduling per cover fails to factor in guest behavior and trends, which impact your employees. The time and effort required to deliver exceptional customer service to different types of tables varies. For example, if guests at one table order soft drinks and appetizers, while a different table opts for a multi-course meal and a few rounds of cocktails, the amount of time required for staff to prepare for each differs, and so does the amount of time needed to serve, clear and tend to each set of guests.
To make this equation even harder to balance, the table with the multi-course meal could have only two covers, while the four-top may be just stopping by for a light post-work snack. Similarly, on a night with low cover counts, the requirements on your staff could still be high, depending on the complexity and volume of orders from any given table.
While this approach has obvious limitations, it is a step in the scientific direction.
The Fourth solution: Demand forecasting
Demand (or item-based) forecasting is currently the most scientific approach to scheduling, as it fully utilizes the power of data. Harnessing the data that exists within the organization (including historical sales analysis and recent trends), combining it with external information (such as weather, national and local events, holidays) and finally overlaying local manager knowledge makes it very possible to predict who will be needed for a given shift.
Self-learning algorithms take all of this data and accurately forecast the individual items that will be sold in 60-, 30- or 15-minute increments. This allows operators to more accurately forecast when they’re likely to sell that $150 champagne and what times will be full of less expensive, non-alcoholic beverage orders, so they can staff accordingly.
Demand forecasting also takes into account exactly how much time is needed to deliver a certain activity, including non-revenue-generating (but necessary) activities like prep and finish. With all of this data factored in, the system then suggests the number of employees needed in each area, at each stage of the day, so the demand can be met. Aside from producing a more accurate forecast the first time, managers can also then stagger shift start and end times to maximize productivity and minimize labor spend.
The benefits abound. Last-second shift swapping decreases, and with a more accurate forecast, managers no longer have to cut employees mid-shift. As a result, your staff have a better sense of what to expect on a given shift, feel more relaxed and prepared to handle the workload, and are better able to drive sales and deliver a winning guest experience. They’re also better able to plan their time and schedule outside of work, which will help them feel more engaged when they’re there.
Conclusion
With labor as the largest line item in most restaurants’ P&L, creating an accurate schedule has always been top-of-mind. Equipped with a scientific approach that uses existing data, trend analysis and external factors, and combining that data with the manager’s local knowledge, managers have the potential for the highest level of accuracy. Matching staffing with workload demand means that they can more confidently get the right people on the right shift at the right time, making for happier employees, more satisfied guests and a healthier bottom line.
To learn more about demand forecasting and how to optimize your scheduling, download Fourth’s complimentary Science of Scheduling white paper, or get in touch here.
This post is sponsored by Fourth